Eclogite Fades Regional Metamorphism of Hydrous Mafic Rocks … · 2017-08-02 · Eclogite Fades Regional Metamorphism of Hydrous Mafic Rocks in the Central Alpine Adula Nappe by

Eclogite Fades Regional Metamorphism of

Hydrous Mafic Rocks in the

Central Alpine Adula Nappe

by CHRISTOPH A. HEINRICH*

Departement fur Erdwissenschaften, Eidgenossische Technische Hochschule,Zurich, CH-8092, Switzerland

{Received 9 November 1984; revised typescript accepted 19 June 1985)

ABSTRACTThe Adula Nappe is a slice of Pre-Mesozoic continental basement affected by Early Alpine

(Mesozoic or Lower Tertiary) high-pressure metamorphism. Mineral compositions in mafic rockscontaining omphacite -I- garnet + quartz record a continuous regional trend of increasing recrystalliza-tion temperatures and pressures that can be ascribed to this regional high-pressure metamorphic event.P-T estimates derived from mineral compositions grade from about 12 kb and 500 °C or less in thenorth of the nappe to more than 20 kb/800 °C in the south.

The regional P-T trend is associated with a mineralogical transition from assemblages containingadditional albite and abundant amphiboles, epidote minerals, and white micas in the north(omphacite-garnet amphibolites) to kyanite eclogites containing smaller amounts of hornblende andzoisite in the south. Textures and mineral compositional data show that these hydrous and anhydroussilicates associated with omphacite + garnet-I-quartz are primary parts of the high-pressure assem-blages. Observed phase relations between these primary silicates, theoretical Schreinemakers analysis,and the thermobarometric results, together indicate that the regional transition from omphaciteamphibolites to kyanite eclogites can be described by two simplified reactions:

alb + epi + hbl = omp + kya -I- qtz + par (H2O-conserving) (15)

par + epi + hbl + qtz = omp + kya + H2O (dehydration) (12)

which have the character of isograd reactions.Local variations of water activity (aH2<3) as indicated by isofacial mineral assemblages, and the

H2O-conserving character of the reaction (15), are interpreted to reflect largely H2O-undersaturatedand predominantly fluid-absent high-pressure metamorphism within the northern part of the nappe.The omphacite amphibolites and paragonite eclogites in this area are thought to have formed byH2O-conserving reactions from Pre-Mesozoic high-grade amphibolites, i.e. from protoliths of similarbulk H2O-content.

The second 'isograd' (12) is interpreted to mark the regional transition from largely fluid-absentmetamorphism in the north to fluid-present metamorphism in the south, where metamorphic pressuresand temperatures in excess of 12-15 kb and 500-600 °C were sufficient for prograde in-situ dehydrationof similar hydrous protoliths to kyanite eclogites. The observation of abundant veins, filled withquartz + kyanite + omphacite, suggests that a free fluid coexisted locally with the kyanite eclogites ofthe southern Adula Nappe at some time during progressive dehydration.

INTRODUCTION

Dehydration of hydrous mafic rocks and associated metasediments to eclogitic rocksduring subduction of oceanic crust is believed to be an important factor in the formation ofpartial melts in the descending slab, the overlying mantle, or the lower crust, giving rise to

• Present address: Bureau of Mineral Resources, Geology and Geophysics, Division of Petrology andGeochemistry, GPO Box 378, Canberra, ACT 2601, Australia.

[Journal ofPttrvtogy, Vd 27, Pwi 1, pp. 123-154, 1986] C Oxford Univenity Pro! 1986

124 C. A. HEINRICH

calc-alkaline orogenic magmatism (e.g. Green & Ringwood, 1968; Wyllie, 1971, p. 208).Evidence for high-pressure dehydration of mafic rocks to eclogites has been established byexperiments concentrating mainly on dehydration-melting relationships at relatively hightemperatures (e.g. Lambert & Wyllie, 1972). However, little is yet known about whichsubsolidus mineral reactions are important under the lower-temperature conditions likely toprevail in the thermally disturbed realm of a subduction or continental collision zone(Anderson et al, 1978; England & Thompson, 1984; Rubie, 1984).

Eclogites currently found in crustal metamorphic terrains often contain assemblagesindicating high pressures and low to moderate temperatures, and many authors havetherefore interpreted eclogite occurrences as evidence for earlier crustal subduction (e.g.Krogh, 1977; Evans & Trommsdorff, 1978; Evans et al., 1979; Ernst & Dal Piaz, 1978;Rubie, 1984). However, it is not clear whether any of these 'crustal' eclogites have indeedformed by dehydration of more hydrous mafic precursors, or whether they have crystallizedunder H2O-undersaturated conditions from primarily anhydrous protoliths. The latterpossibility was first suggested by Bearth (1959) for nearly anhydrous eclogitic pillowmetabasalts embedded in glaucophane-rich metahyaloclastites in the Western Alps. Theconcept that eclogites found in metamorphic terrains represent the H2O-undersaturatedequivalents of blueschist or amphibolite assemblages, forming at the same P-T conditionsbut higher H2O activity, has been supported more generally by Fry & Fyfe (1969), Newton &Fyfe (1976) and in many textbooks (e.g. Miyashiro, 1973, p. 318; Winkler, 1974, p. 274). Otherauthors have stressed the observation of eclogitic veins and primary hydrous silicates asevidence for the presence of a fluid and high H2O activity during the formation of someeclogites (Essene & Fyfe, 1967; Holland, 1979a; Brown & Bradshaw, 1979).

Eclogites from various terrains are mineralogically diverse (Smulikowski, 1964; Colemanet al., 1965) and have probably formed under a wide range of P- T-aHlO conditions. However,comparison between eclogites and a more general interpretation of their genesis is hamperedby the structural incoherence typical of most terrains, which rarely show regular regionaldistribution patterns of eclogitic assemblages. The few known exceptions are from SWNorway (Krogh, 1977; Griffin et al, 1984), New Caledonia (Black, 1977), the Adula Nappe(Heinrich, 1982), and from Venezuela, where Maresch (1983) first reported a field exampleindicating progressive in-situ dehydration of amphibolite to eclogite. Recently, Takasu(1984) presented textural evidence for the transformation of epidote amphibolite to hydrouseclogite, which he ascribed to prograde dehydration metamorphism.

The Central Alpine Adula Nappe (Switzerland) displays a regionally systematic sequenceof high-pressure assemblage in mafic rocks, grading from hydrosilicate-rich garnet +omphacite + quartz-bearing amphibolites, blueschists and eclogites in the north, to quartz-kyanite eclogites containing only minor hornblende and zoisite in the south (Heinrich, 1982,1983). The southernmost occurrences are associated with garnet metaperidotites (Ernst,1977, 1978; Evans & Trommsdorff, 1978; Evans et al., 1979). This suite of hydrous maficrocks has been chosen for a petrological study, to define some of the mineral reactions thatare likely to control the formation of different omphacite + garnet + quartz-bearing assem-blages, and to place some constraints on the conditions under which hydrous maficprotoliths may undergo prograde dehydration to hydrate-poor eclogites.

REGIONAL GEOLOGY AND STRUCTURAL RELATIONS

Structure and lithology of Adula Nappe

The Adula Nappe is an east- to southward dipping sheet-like body with an estimatedthickness of 2-4 km, continuously exposed along a north-south profile of 50 km. It consists

ECLOGITE METAMORPHISM IN THE ADULA NAPPE 12S

10 kmVALS

Olivone • Hinterrhein

Son Bernardino

CONFIN

TRESCOLMEN

DURIA

• Bellinzona

Lithological units of Adula Nappe basement

| metagranitoid gneisses predominant;-.;-,] (semi-)pelitic schists and gneisses with mafic lenses predominant

Mineral assemblages in quartz - bearing mafic rocks

A albite + hornblende + epidote ( tomp, white micas)A omphacite + epidote + porogonite (± amp, phengite)

§ kyonite + omphacite + paragonite (± amp, phengite)

kyonite + omphacite (± amp, epi, white micas)FIG. 1. Geological map showing the outline and internal structure of the Adula Nappe in southern Switzerland(simplified from Heinnch, 1983, plate 1). The metascdimentary units of the Adula Nappe host mafic lensescontaining high-pressure assemblages. Occurrences of diagnostic quartz-bearing assemblages are indicated by

specific symbols, including some data by Baumgartner (1981).

predominantly of metagranitic and metapelitic gneisses and schists of a Pre-Mesozoiccontinental basement, which has been affected first by Mesozoic or lower Tertiaryhigh-pressure metamorphism, followed by Oligocene Barrovian-style metamorphism. Thetwo main rock types are interlayered on a kilometre scale grossly parallel to the nappestructure (Fig. 1). High-pressure assemblages are mainly preserved in mafic lenses within themetasedimentary layers (Heinrich, 1982; including additional references).

126 C. A. HEINRICH

Carboniferous Rb/Sr age data indicating an event of Pre-Alpine high-grade meta-morphism are preserved in some white micas from the northern Adula Nappe (Frey et al,1976). Rb/Sr whole rock data suggest a Variscan (Late Carboniferous) origin for migmatitesin the southern Adula Nappe, with local remobilization during Tertiary amphibolite faciesmetamorphism (Hanny et al, 1975). Some mafic high-pressure assemblages in the northernpart of the nappe can be shown to derive from earlier high-grade amphibolites on the basis oflocally preserved relict textures, in which glaucophane rims and sphene precipitates havepartly replaced coarse-grained, formerly Ti-rich hornblendes with an earlier parallelalignment (Plas, 1959, p. 526; Heinrich, 1983, p. 78). Since the metasedimentary layersenclosing the mafic rocks extend over most of the length of the nappe, a similar amphiboliticorigin is likely for the majority of the eclogitic rocks described in this paper. As in mafic rocksfrom other units of Variscan continental basement in the Alps, which are less affected byAlpine metamorphism (e.g. Buletti, 1983), trace element geochemistry indicates an ultimateorigin from oceanic basalts for at least some of the Adula (Evans et al., 1979, 1981;Aurisicchio et al, 1982).

Mesozoic carbonate-rich to pelitic metasediments and minor metabasites separate theAdula Nappe from over- and underlying basement nappes. Some of this material has beenimbricated at least into the frontal (northern) part of the Adula Nappe prior to all recognizedstages of Alpine metamorphism and deformation (Jenny et al, 1923; Plas, 1959; Egli, 1966;Baumgartner & Low, 1983). Nevertheless, Mesozoic rocks comprise at most a few volumeper cent of the present Adula Nappe.

The time of eclogite facies metamorphism has not been dated radiometrically, butgeological evidence constrains it to the interval between Late Jurassic and Early Tertiary(Heinrich, 1982; 1983, p. 175).

Post-eclogite facies deformation and overprinting

Rocks containing garnet + omphacite + quartz (referred to as eclogitic assemblagessensu lato) are distributed over most of the Adula Nappe. They usually occur as relictsin cores of mafic lenses, where they escaped later deformation and hydration to lower-pressure greenschist and amphibolite assemblages. Heinrich (1982) described some keyexposures which show that eclogite facies metamorphism was shared by the (semi-)peliticcountry rocks. In most places, however, high-pressure assemblages were preserved inmafic rocks only. The phengite- and paragonite-rich high-pressure assemblages of thepelitic rocks were usually completely overprinted by prograde, biotite + feldspar-formingdehydration reactions, which occurred during later amphibolite facies metamorphism(Heinrich, 1982).

Post-eclogite deformation of schists and gneisses included isoclinal folding which reachesthe scale of kilometres in the southern part of the nappe (Jenny et al, 1923; Egli, 1966;Codoni, 1981). This deformation probably occurred before and during nappe emplacementand Oligocene Barrovian-type metamorphism, but after regional eclogite facies meta-morphism. No original length scale can therefore be attributed to the still regular, butcertainly distorted and blurred, exposed distribution pattern of eclogitic assemblages.

PETROGRAPHY OF ECLOGITIC ASSEMBLAGES

Mineral textures in thin section, combined with structural observations in zoned maficlenses, allow a clear distinction of assemblages produced by post-eclogite overprinting fromearlier eclogite facies assemblages. Overprinting to amphibolite or greenschist invariablyproceeded via a texturally distinctive intermediate stage, involving symplectite pseudo-

ECLOGITE METAMORPHISM IN THE ADULA NAPPE 127

TABLE 1

Primary assemblages and visually estimated modal composition of rocks studied by electronmicroprobe

ValsAd42-9-14*VA30Ad41-9-20Ad65-O-8*

ConfinAd48-9-5*E5-10E5-6Ad51-9-12Ad51-9-17

TrescolmenAd2 5-9-3*Ad61-9-l*

GagnontCH271*Mgl63KMgl63-*-8E2-2

AramlMg9-5-12c«

gar

15205

10

35203510

3020

35353520

40

cpx

152020

5

3030354520

4015

35353530

45

kya

5

1010

10

5

qt:

5X

10X

51010X

X

X

30

101520X

X

amp

20 hbl40 hbl

5 hbl60glc

15 hbl20 bar

5 hbl15 hW20 act

10 hbl15glc

10 hbl10 hbl20 hbl30 hbl

5 hbl

ZOl

15

2515

X

X

205

5

clz-epi

10X

10

105

10

par

5

55

5

phe

15X

55

X

10

5

additional phases

20 alb; pyr, rutpyr, rut; clcpyr; sph; grapyr, sph

pyr; rutpyr; rutrutpyr, nit; sph50 dol; dc; pyr; sph

rut5 tic; pyr; rut

rut; iLmrot; llmrutrut

rot; llm

Numbers refer to volume per cent, x = 2-5 per cent * denotes samples for which silicate analyses are given in Tables 2 and 3Exact sampling localities and some petrographic observations are listed in Appendix 1.

The following abbreviations for mineral phases are used throughout act = actinolite, alb = albite, amp = amphibole,bar = barroisite, clc = calcite, dz = clinozosile, cpx = clinopyroxenc, dol = dolomite, epi = epidole, gar — garnet, gle — glauco-phane, gra = graphite, hbl = hornblende, llm = ilmenite, jad — jadeite, kya = kyanite, mus = muscovite, omp — omphacite,par = paragonite, phe = phengite, pyr = pyroxene, pig - pl»giodase, qtz = quartz, rut = ralile, sph = sphene, tk = talc, zoi =zoisite.

morphs and thin reaction rims along grain boundaries (Heinrich, 1982). By contrast, theminerals here considered to be part of a 'primary eclogitic assemblage', which are the subjectof this study, are characterized texturally by sharp mutual grain boundaries, and by grainscomparable or larger in size than omphacite and quartz (Table 1, Appendix 1, Fig. 2). Manyof the assemblages so defined show equigranular textures with or without a linear or planarparallel orientation shared by a variety of anhydrous and hydrous silicates. Some mineralscan occur as euhedral (garnet) or anhedral-poikilitic (hornblende, kyanite, primary albite)porphyroblasts. Amphibole poikiloblasts appear to overgrow the aligned fabric in thematrix of some eclogites, but mutual inclusions involving minerals of similar composition,such as hornblende + garnet + omphacite + quartz inclusions in kyanite next to kyanite +garnet + omphacite + quartz inclusions in hornblende (e.g. CH271, Ad25-9-3), indicate alargely contemporaneous crystallization of hydrous and anhydrous 'primary' silicates. Eventhe highest-grade kyanite eclogites (see below) from the Arami eclogite-garnet meta-peridotite complex contain zoisite and hornblende in parallel alignment with equigranulargarnet + omphacite + kyanite + quartz, and homogeneous garnet grains contain inclusionsof hornblende of the same composition as the matrix. Despite extensive overprinting bytexturally later and chemically different amphibole assemblages, there is little evidence tosuggest (cf. Ernst, 1977, p. 377) that all hydrous phases formed at the expense of an earlier,completely anhydrous eclogite assemblage.

In the porphyroblastic eclogite textures, chemical equilibrium was probably approachedat least at contacting grain boundaries, but rarely attained throughout a whole rock even at

C. A HEINRICH

JOOym 1

PH- J q t Z

alb

alb cpx , y

FIG. 2. Typical mineral textures of primary high-pressure assemblages, (left) Amphibolite Ad42-9-14 from Vals,with coexisting omphacite + albite +quartz + garnet + hydrous silicates, (right) Quartz-kyanite ecloghe Ad25-9-3

from Trescolmen, with primary hornblende, zoisite and phengite. See Table 1 for mineral abbreviations.

the scale of a thin section. This is shown by the chemical zoning of larger mineral grains,indicated by arrows in Figs. 3, 8, 9 and 10, and discussed below. Deviation from completeequilibration is particularly obvious in chemically and texturally zoned garnet porphyro-blasts, which in their Fe- and Ca-enriched core contain inclusions of biotite, sphene andrarely plagioclase (phases that do not occur in the eclogite matrix) and of dark greenamphibole, which has a highly aluminous composition different from the pale bluish tobrown matrix amphiboles. These inclusions could represent armoured relics of Pre-Alpineamphibolite assemblages, but the presence of omphacite and epidote as additional inclusionminerals indicates that the inclusion assemblages probably record an early (omphacite-epidote amphibolite?) stage in the Alpine high-pressure metamorphic history of eclogiteformation (see below).

Many kyanite eclogites, especially in the Trescolmen area (Fig. 1), contain irregular veins10-100 mm in width and less than a metre in length. The veins are filled with quartz, minoromphacite, sometimes apatite, and euhedral kyanite crystals up to several centimetres in size.In many cases they cut eclogite schistosity and banding. Vein walls show no obviousmineralogical zonation or alteration of the adjacent host eclogite. All matrix minerals of thehost eclogite occur at the vein contact, with omphacite sometimes growing as needles intothe vein. Although quartz has recrystallized to polygonal grains, the textural relationsindicate that all vein minerals crystallized from a vein fluid which was probably in localequilibrium with the immediately adjacent kyanite eclogite.

Quartz and kyanite contain fluid inclusions up to 20 /mi in size. Most inclusions areH2O-rich with a small gas bubble, but highly saline inclusions with cubic ?NaCl daughtercrystals have also been observed. No microthermometric studies have been performed so far,because most fluid inclusions are clearly secondary. At least the high-salinity inclusionshave probably formed during amphibolite fades overprinting. This is indicated by theirarrangement along healed fractures in the extension of hairline veins in the adjacent eclogite,along which omphadte and garnet are replaced by dark green hornblende, diopside, andplagioclase. These later veins or joints occur even in the freshest eclogites and represent aninitial stage of post-eclogite facies overprinting controlled by the access of external fluids(cf. Heinrich, 1982, p. 33).

From a total of about 200 samples studied in thin section, 2 to 6 omphacite + quartz-bearing rocks from 4 areas of less than 1 km2 each (Gagnone, Trescolmen, Confin and Vals;Fig. 1), plus a few additional samples from other localities in the Adula Nappe, were chosenfor electron microprobe investigation. They were selected on the basis of their showingrelatively simple textures among 'primary' hydrous and anhydrous silicates, with no or

E C L O G I T E M E T A M O R P H I S M IN T H E ADULA N A P P E

Mg

A Gagnone, Arami

O Confin.Trescolmen

• Vals

129

30

Fe + Mn

di + hd + Cats

FIG. 3. (a, above) Composition of eclogitic garnets, with arrows indicating trends from core to ran of zonedporphyroblasts. Empirical boundaries for mole per cent pyrope (py30 and py}5) separating type-A, -B, and -Ceclogites are from Coleman el ai. (1965). (b, below) Composition of Adula clinopyroxenes in the triangle

di + hd + CaAI2SiO6 (Cats) = Ca, jd = Na-Fe3 + , and ac = Fe3 + , after Essene & Fyfe (1967).

only traces of incipient retrogression to amphibolitic symplectite textures (Table 1).Representative point analyses are given in Tables 2 and 3 together with analytical andrecalculation procedures.

P-TESTIMATES FROM MINERAL CHEMISTRY OFGARNETS AND PYROXENES

The composition of the anhydrous eclogite minerals, garnet and clinopyroxene (Fig. 3),constrains the pressure-temperature conditions of the formation of quartz-bearing eclogites,independent of the presence of additional hydrous phases.

130 C. A. HEINRICH

Fe2 + /Mg fractionation between garnet and clinopyroxene

The temperature-dependent exchange equilibrium

(1)

(vector notation after Thompson, 1981; Thompson et ai, 1982) is relatively pressureinsensitive, and can be used to estimate the formation temperature of eclogitic assemblages.Fig. 4a is a log-log plot showing the Fe2+/Mg distribution data from the Adula Nappeaccording to equilibrium (1). Contours for constant distribution coefficients KD = 65 andKD = 17, where

Fe

separate empirical KD ranges for garnet-omphacite pairs from different geological settings assuggested by Banno & Matsui (1965) and Lovering & White (1969). Largely because ofchanging garnet compositions (Fig. 3a), KD values from the northern and southern part ofthe Adula Nappe plot in clearly different fields, in general agreement with their presentmetamorphic environment. Ellis & Green (1979) have calibrated equilibrium (1) as afunction of temperature, pressure, and the Ca-content of the system, expressed by thegrossular content (Xf?,r). Their equation (9) can be rearranged to

_Fe_Var_ll48 / FeV p i /1316 4-72 FN

The left hand term of equation (2) may formally be considered as a garnet composition term,empirically corrected for the non-ideality in garnet. A first approximation for T of 773, 873,973 and 1073 K for Vals, Confin, Trescolmen and Gagnone/Duria/Arami was used for thecorrection, which is only weakly dependent on the T values. The garnet composition datathus corrected are shown as a function of log(Fe2+/Mg)cpi in Fig. 4b. The 'Ca-correcteddistribution coefficient', Kg, is plotted in Figs. 4b, 5 and as a function of temperature andpressure in Fig. 6, where:

l o g K g = l o g K D - ^ J r £ r . (3)

Fig. 4b shows that Ellis & Green's calibration corrects for all compositional effects amongthe samples from the northern and southern Adula localities, indicating a trend in Kg suchthat:

Kg(Arami,Duria) < Kg(Gagnone) < Kg (Vals).

The distribution coefficients for the intermediate Confin and Trescolmen localities varyconsiderably even after correction for Xf£. Kg tends to decrease with increasing jadeitecontent of the particularly sodic omphacites from each of these two localities (Fig. 5). Asimilar effect has been found for jadeitic pyroxenes of the Western Alps, where Koons (1984)proposed preferential ordering of Fe relative to Mg into the M2-site of clinopyroxene (solidsolution towards clinoferrosilite) as a likely cause for this effect. Koons (1984) suggested that,as the total Fe + Mg in Ml is high (as in the jadeite-poor pyroxenes used in Ellis & Green'sexperiments), ordering of relatively minor Fe in M2 has no strong effect on the bulk Fe/Mgratio of the pyroxene. However, the preference of Fe2 + for the slightly larger M2 site becomesrelatively more important with increasing jadeite content (and hence decreasing totalFe + Mg content) of very sodic omphacites. The shape of the dashed curves in Fig. 5 iscalculated to represent the maximum estimated effect of M1, M2 ordering on Kg. They were


10

0.5

0.0

-0.5

0 5

-0.5

|log(Fe/Mg)gar

•oAAO

Vals

ConfinTrescolmen

Gagnone

Arami, Duria

c

'glaucophane schistterrains"

"amphibollte-^granulite -

terrains/ ' ' "kimberlitepipes"

KD=17 D 6.5 A

log(Fe2+/Mg)cpx

2+ CDX

log(Fe /Mg)p

-1.5 -1.0 -0.5FIG 4. (a, above) Logarithmic distribution diagram of Fe2 + /Mg between garnets and clinopyroxenes from theAdula Nappe. Fe2+(cpx) is estimated from cation normalized stoichiometry (see Table 2), while for garnetFe(gar) = Fe"" was plotted neglecting minor fernc iron. Empirical boundaries of KD = 6-5 and 17 separatingtype-A, -B and -C cclogites are from Lovering & White (1969). (b, below) The same data as in (a\ except thatthe garnet composition term (vertical axis) is corrected for Xf£ after Ellis & Green (1979) (eqn. (2\ text). TheCa-corrected distribution coefficient Kg, indicated by the 45° contours, is constant for the mineral pairs from eachof the localities Vals, Gagnone and Arami/Duria, but widely variable for pairs from Trescolmen and Confin

(see text and Fig. 5).

132 C. A. HEINRICH

TABLE 2

Electron microprobe analyses of clinopyroxenes and garnets

Chnopyroxene

SlO2

TiO2

A12O3

FeO(tot)MnOMgOCaONa 2 O

Total

Cations normalized toSiTiAl(tet)Al(oct)CrFe3 +

Fe2 +

MnMgCaNa

Garnet

SiO2

TiO2

A12O3

Cr 2 O 3

FeO(tot)MnOMgOCaO

Total

Cations norn

SiTiAlCrFeMnMgCa

Ad42-9-14

55-17022

1113OOO4-990088O0

13-246-81

99-65

£<cations) = 6

1-97001003044OOO005010OOO043051047

Ad42-9-l4 Ad65-0-8rim

38-65Oi l

21-85OOO

27-350663-359-81

101-79

lalized to

2-990011 99OOO1-77OO4039O8I

run

37 55012

2099OOO

27-640961 54

1097

99 77

Locations) = 8

3000011-97OOO1-85007018094

Ad65-0S

55-20012

11-61OOO8O10055-43

10158-53

99O9

1-99OOO001048OOO013012OOO029039060

Ad48-9-5run

38 761 10

21 81006

21410267-838-67

99-90

2-960061-97OOO1 37O02089071

Ad48-9-5

55-80OOO

11-120052-81OOO8-91

13-556-86

9910

1-99OOOOOI0-45OOO003005OOO0-470-520-47

Ad48-9-5core

38 61009

21 66OOO

21O41 50529

10-70

98 87

3 010-011 99OOO1-370-100-62O89

Ad25-9-3

5514007

IO08O072-41OOO

1O4315-605-47

99-27

1-97OOO003039000001006OOO056060038

Ad25-9-3rim

4O05004

22-56000

19-39043

11-176-90

10O55

2-99OOO1-99OOO1-21003124055

Ad6l-9-l

57-93006

17850032 62OOO4 79684

1072

10O84

199OOO001071OOO001007OOO025025071

Ad61-9-lrun

39 29003

22 82OOO

22 95OOO

1025507

10O4I

2-97OOO2O3OOO1-45OOO115041

CH271run

39-12OOO

22 25OOO

22 870519-895 88

I0O52

2 97OOO1-99OOO1450031 12048

CH27I

56-10006

1262OOO2 59OOO8 38

13077 23

10O04

1-97OOO003050OOO002006OOO044049049

CH271core

39-51015

22O0OOO

21-400539OI8-31

I0O92

2-99OOI196OOO1 35003101067

Mg9-5-12c

55-33012

11-36OOO3O70049-27

14-446-33

99-97

1-96OOO004043OOO004005OOO049055044

Mg9-5-12crim

39-19006

21-89000

18-530477-81

11-94

99 90

299OOO1-98000118003089098

Single representative point analyses of contacting mineral grams are listed above and in Table 3. Unless indicated otherwise, rimor plateau compositions are given See Table 1 and Appendix 1 for sample description

Analytical conditions: Wave-length dispersive spectrometers on automated ARL-SEMQ microprobe; 15 kV acceleration voltage,20 nA reference sample current on brass, 20 sec counting time; 10 //m defocujsed beam for micas to prevent alkali loss. Simple naturalsilicate and oxide standards. ZAF-correction, including fixed amounts of H2O for hydrosilicates (Table 3), using program EMMA(J Sommcraucr and R. Gubser, ETH Zurich). Normalization of clinopyroxene analyses to 6 cations (Laird & Albee, 1981,appendix I) nearly always resulted in fully charge-balanced formulae of ideal site occupancy with slightly less than 200 Si per6 cations, thus illowing unique assignment of Fe2 + aiid Fe3 +

calculated for (in atoms per 6 cations): Fe2 +(M2) = 004, Mg(M2) = 0, Fe2 +/Mg(M 1) = 0-1.Assuming maximum possible site preference, these values could be representative for theanalysed Na-rich omphacites (Ad61-9-l, E5-10; Table 2), indicating that the orderingmechanism proposed by Koons (1984) has the right magnitude to explain the compositional


1.2

0.8

0.4

flog KQ (gar/cpx)

- -f-

Trescolmen

.Na c p x

Ql 0.2 0.3 0.4 0.5 06 0.7

FIG. 5. The compositional dependence of the garnet-clinopyroxene Fe/Mg distribution coefficient remaining afterapplication of the correction for Kg by Ellis & Green (1979). Log Kg (see eqn. (3), text) tends to decrease withincreasing Na-content of the relatively jadeite-rich omphacites from each of the areas Confin (circles) andTrescolmen (tnangles). The dashed lines (at arbitrary vertical position) are theoretically calculated to show themaximum effect which preferential Fe 2 + ordering in the clinopyroxene M2-site is expected to have on log Kf,

((Coons, 1984)

dependence of log K& observed in the Adula assemblages. Unfortunately the analyticalobservations scatter too much to place any quantitative limits on this ordering model.

Excluding omphacites with Xjd > 06, where cation ordering is more likely to causesystematic errors in temperature estimation, the largest uncertainty in applying Ellis &Green's thermometer to Adula eclogites is probably mineral inhomogeneity. Error bars inFigs. 4 and 5 represent the total variation of 5-20 point analyses of several adjacent garnetand pyroxene grains, excluding only the Ca-Fe-rich compositions of inclusion-rich cores oftexturally zoned garnet porphyroblasts. The apparent uncertainty could have been reducedby selecting only analyses of rims in immediate contact, a procedure which generally yieldsthe lowest values for Kj> However, such a selection could introduce a systematic error,because of the possibility of local post-eclogite Fe/Mg exchange along grain boundariesduring amphibolite facies overprinting. With the wider error bars shown in Fig. 4, the totaluncertainty due to incomplete chemical equilibrium probably outweighs the errors resultingfrom stoichiometric estimation of Fe 2 + /Fe 3 + in pyroxene, because normalization to 6cations allowed a unique assignment of cation occupancies to an ideal charge-balancedpyroxene stoichiometry in most cases. Possible exceptions are some clinopyroxenes withvery low total Fe/Mg, notably ultramafic garnet peridotites where uncertainties due tostoichiometric Fe 2 + /Fe 3 + estimation may be larger than in mafic eclogites (e.g. the point inthe lower left corner of Fig. 4 representing the analyses from the Gagnone peridotite givenby Evans & Trommsdorff, 1978).

Jadeite content (Xj(i) of clinopyroxene coexisting with quartz

In combination with Fe-Mg exchange thermometry, the net-transfer equilibrium

NaAlSi2O6(cpx) = NaAlSi3O8(plg)-l-SiOj(qtz) (4)

places limits on the formation pressure of omphacite +quartz ±albite, but uncertainties dueto ordering in omphacite restrict its application to natural assemblages.

Site ordering of Na,Al and Ca,Mg + Fe causes preferential stabilization of clinopyroxeneswith compositions close to jadeite: diopside + hedenbergite = 1:1, which leads to twomiscibility gaps in the system NaAlSi2O6-Ca(Mg,Fe)Si2O6 at low temperatures (Clark &Papike, 1968; Carpenter, 1980). Ordering can occur during cooling of originally disorderedhigh-temperature omphacites. Alternatively it may be primary in lower-temperatureomphacites from blueschist terrains, in which case it is often reflected by gaps in

134 C A. HE1NRICH

10 -

400 500 600 7 0 0 800 9 0 0

FIG. 6. P T locations of some experimentally determined equilibria, and likely conditions of crystallization of thequartz + omphacite + garnet assemblages (shaded areas) from Vals (V), Confin (C), Trescolmen (T), Gagnone (G)and Arami/Duria (A). Contours of ACg refer to the Ca-corrected distribution coefficient of Fe2 +/Mg between garnetand clinopyroxene (eqn. (3), text), after Ellis & Green (1979). Isopleths of jadeite (jd, mole per cent) in omphaciteaccording to equilibrium (4) after Gasparik & Lindsley (1980). The preferred stabilization of ordered P2/nomphacites to pressures below the dashed jd50 contour (P2/n-C2/c boundary) is schematic only (cf. Holland, 1983).Dashed contours 1-7, 1-8 and 21 are for the Al-contents of orthopyroxenes in garnet lherzolites from Gagnone,Arami and Duna, respectively, from data given by Evans & Trommsdorff (1978), recalculated using the newcalibration of Harley & Green (1982). Other curves (P H l 0 = P) are from Chatterjee & Johannes (1974) formus + qtz = Al2SiO5 + H2O, from Thompson & Tracy (1979) for the minimum melting of muscovite + quartz, fromChatterjee (1972) and Holland (19796) for the reactions involving paragonite, and from Holdaway (1971) for the

stability field of kyanite.

compositional plots such as Fig. 3b (cf. Carpenter, 1980, fig. 1b). Completely orderedomphacites have not been synthesized in laboratory experiments, but Holland (1983)measured at 600 °C the effect of partial ordering on extending the stability field ofintermediate omphacite compositions ( +quartz) relative to albite. Gasparik & Lindsley(1980) used part of these data to fit the P-Tdependence of the mole fraction of NaAlSi2O6

(Xjd) in disordered and partly ordered clinopyroxenes coexisting with albite + quartz.Contours in Fig. 6 are derived from Gasparik & Lindsley (1980, p. 332).


The omphacites from Confin, Trescolmen and Gagnone are acmite-poor and cover thewhole compositional range from diopside-hedenbergite to more than 70 mol per centjadeite (Fig. 3b). The absence of any compositional gaps suggests that at least thoseclinopyroxenes with compositions outside Xjd = 0-5 ±0-1 have originally crystallized in thedisordered C2/c state (cf. Carpenter, 1980, fig. la). This suggestion is in agreement with therelatively higher equilibration temperatures indicated for the southern parts of the nappe bythe Fe/Mg distribution data. The contours for equation (4) according to Gasparik andLindsley (1980) therefore allow a realistic estimation of the minimum pressures required tostabilize the plagioclase-free eclogites from Confin, Trescolmen and the southern AdulaNappe (Fig. 6).

By contrast, all omphacites from the northern Adula Vals locality plot in a smallwedge-shaped field in Fig. 3b (dashed lines) with its apex near Xjd = 0-5. Comparison withfig. 1b of Carpenter (1980) suggests primary crystallization of the Vals pyroxenes in theordered P2/n omphacite structure, in agreement with a relatively low crystallizationtemperature. Because of the unknown effect of complete ordering on stabilizing low-temperature omphacites, the extrapolated jd J 0 isopleth (Fig. 6) after Gasparik & Lindsleyprobably provides an upper limit for the formation of the omphacite + quartz + albiteassemblages from Vals. The true crystallization pressure may be a few kilobars lower, assketched in Fig. 6 by the qualitative curve jd(omp) + qtz = alb and the shaded area 'V. Anyerrors in pressure estimation due to ordering in omphacite would be accentuated if extra-polation of Ellis & Green's Fe/Mg geothermometer caused any over-estimation of crystalliza-tion temperatures due to ordering or other non-ideality effects in omphacite and garnet solidsolutions. The P-T limits for the Vals area are therefore subject to the largest uncertainty.

Al-content of orthopyroxene

Eclogite formation pressures can be further constrained from published Al-contents oforthopyroxenes in garnet Iherzolites from Gagnone, Arami and Duria, if it is assumed thatthese rocks formed at the same metamorphic conditions as the associated eclogites. Thisassumption is debated (Ernst, 1978; Evans & Trommsdorff, 1978) but compatible with thedata of Evans et al. (1979) and the present eclogite data (Fig. 6). Orthopyroxene from a singleoccurrence of 2-pyroxene eclogite at Trescolmen has an Al-content below 0-01 atoms per4 cations, which indicates a lower temperature than for the crystallization of the southernAdula peridotites, but which is too low to sensitively constrain equilibration pressure(Harley & Green, 1982).

Conclusions from thermobarometry

The estimation of absolute values delimiting crystallization conditions of the Adulaeclogites involves errors which may be systematic and which are probably larger thancommonly assumed. Despite this, the mineral compositions of anhydrous eclogite matrixminerals show a systematic trend of regional increase in recorded temperatures and pressuresfrom north to south through the present Adula Nappe. Combining all available data, thefollowing values are considered to be the most likely P-T brackets for the matrixrecrystallization of eclogitic rocks from different areas of the Adula Nappe:

North

South

ValsConfinTrescolmenGagnoneArami, Duria

450-550450-550550-650600-700750-900

°C°C°C°C°C

10-1312-2215-2215-2518-35

kbkbkbkbkb

TABLE 3

Electron microprobe analyses of primary hydrous silicates and albite

Amphibole Ad42-9-14 Ad65-0S Ad48-9-5 Ad25-9-3 Ad61-9-1 CH27I Mg9-5-12c

SiO2

TiO2

A12O3

Cr2O3

FeO(tot)MnOMgOCaONa2OK2OH 2 O *

45-04061

14120-00

12-26O09

11-328-764-020-492-50

54-67008

11-18OOO

13-690-088 912-156-20O052-50

47-700-27

12-750-058-26OOO

14-709013-410-322 50

49-42023

12-430-055-68

OOO16-268-903-34O322-50

58-21O10

13-82OOO3-65OOO

13-591-367060-152-50

4<5-290-56

16-35OOO6-50

00414-307-205-200-752-50

43O8089

15-640078-04006

14-6710733-740O62-50

O>

X

zO•T"

Tola! 99-21 99-52

Cations normalized to E(cations) — Ca — Na — K. = 13Si 6-55 7-67Ti 0O7 O01Al(tet) 1-45 033Al(oct) 097 1-51Cr OOO OOOF e 3 + 039 046Fe2+ 111 1-15Mn 001 O01Mg 2-46 1-86Ca 1-36 O32Na(M4) 064 1-68Na(A) 050 OOIK 0O9 OOI

98-97 9913 100 44 99-67 99-46

6-760O31-24088OOI0570410003101-37063O30006

6-89O02111093OOI051015OOO3-381-33067023006

7-730010271 89OOO012028OOO2-690191 -81OOI003

6-49006151119OOO050026OOO2-99108092049013

6160091-84080OOI050046OOO3131-65036068OOI

Albite, epidote-yroup minerals and sheet

SiO2

TiOjA12O3

Cr2O3

FeO(tot)tMnOMgOCaONajOKjOH2O*

Total

CationsSiTiAlCrFeMnMgCaNaK

albAA42-9-14

68-210-00

20-01OOO0150-00000078

11-35004OOO

100-54

2-97OOO103000001OOOOOO004096OOO

clzAd42-9-14

38 79014

28-21OOO7-35OOO014

23-46OOOOOO1-80

99-89

2-990012-57OOO047OOO0021-94OOOOOO

silicatesepi

Ad65-0S

37-68Oil

2616OOO9-65OilOOO

23-61000OOO180

9913

2-950012-42OOO063001OOO1 98OOO000

clzAd48-9-5

39-18025

29-610034-59000012

24-62OOO0021-80

10O21

2-99OOO2-67OOOO29OOO001202OOOOOO

ZOI

Ad25-9-3

39-84005

32-35OOO1O9005009

23-87004OOO1-80

99-18

3-04OOO2-91OOO007OOO0011-95001OOO

zoiMg9-5-12c

39-24006

32-710041-27OOO009

23-84003OOO1-80

99 07

300OOO2 95OOO008OOO0011 95000OOO

pheAd42-9-14

4917052

29-52OOO2-39OOO2-910011-319-444-50

99-76

3-250032-30OOO013OOO029OOO017080

pheAd25-9-S

51-83033

26-21OOO098OOO4-63002045

10214-50

9916

3-430022O4OOO005OOO046OOO006086

pheAd48-9-5

4917041

28-880051 32OOO3180051029-654-50

98-22

3-30O022-29OOO007OOO032OOO013083

parAd48-9-5

45-45013

38-82003O46OOO0240 536-97O984-50

98O9

2-960012-98OOO003OOO002004088008

parAd61-9-l

46-61Oil

39-81OOOOOOOOO017OOO7-73O184-50

9910

2-98001300OOO000OOO002OOO096002

talcAd61-9-l

6114OOO055OOO6-95OOO

26-95OOO0090024-50

100 20

3-96OOO004OOO038OOO2-61OOO001OOO

nr-uCt

Hm

- i>2

o"0X

2_zHXm>O

cRepresentative analyses of grains in immediate contact with garnet + omphacite (for analyses see Table 2) +quartz ±kyanite (Table 1; Appendix 1) are given. Amphibole

analyses were normalized to a cation sum £(cations) — Ca — Na — K = 13. Site allotment and estimation of Fe 2 + /Fe 3 + using the assumption ( M & F e J ^ ^ O (noanthophyllite-type exchange) mostly resulted in charge-balanced amphibole formulae, in contrast to other possible assumptions (normalization 3 of Laird & Albee, 1981,Appendix 1). The other analyses were normalized to locations) = 5 for albite; locations) = 8 for epidote-group minerals, epidote, clinozoisite, and zoisite; locations) -Ca —Na —K -= 6 for the sheet silicates, phengite, paragonite, and talc.

* Assumed for ZAF correction. f As Fe2O3(tot) in epidote-group mineral analyses.

-o-om

138 C A. H E I N R I C H

In addition to this regional trend, incomplete equilibration as recorded by garnet zoningfrom Fe-rich cores to Mg-rich rims (Fig. 3a) indicates a segment of increasing temperaturealong the P-T path towards crystallization of the matrix assemblage of individual eclogites.

MINERAL CHEMISTRY OF PRIMARY ECLOGITIC HYDROSILICATES

Paragonite has been identified by microprobe in eclogites of the northern and centralAdula Nappe, usually occurring together with phengite. Its composition is displaced onlyslightly from the end-member NaAl3Si3O10(OH)2 by about 01 atom units of KNa_,towards muscovite and an equal displacement along the (negative) Tschermak's vector— Al2(Mg,Fe)_1Si_1. It is often associated with relatively jadeitic omphacites, and itseven-grained textural alignment with omphacite clearly shows it to be a primary eclogitichydrosilicate. The occurrence of coexisting paragonite + omphacite(Ar

Jd = 0-6-0-7) +kyanite according to the equilibrium

NaAl3Si3Oi0(OH)2 = NaAlSi2O6(cpx) + Al2SiO5 + H2O (5)

provides an upper limit for the load pressure of eclogite formation in the Confin andTrescolmen areas (Fig. 6; Holland, 1979b), irrespective of the presence of any fluid phase.

Phengite, the only potassic mica found as primary constituent in Adula eclogites, isdisplaced from muscovite towards celadonite by 0-35-0-55 units of — Al2(Mg,Fe)_!Si_,.This exchange is inversely correlated with a 0-2 to 0-1 unit exchange of KNa_ ( towardsparagonite. The Fe2 + Mg_, exchange equilibrium between phengite and garnet has beenstudied experimentally by Krogh & Raheim (1978) and by Green & Hellman (1982) underreducing conditions where Fe 2 + = Feto1 could be assumed. Natural phengites allowconsiderable exchange along Fe3 + Al _,, but since some displacement towards tri-octahedralmicas is possible as well, no unambiguous estimation of Fe 2 + /Fe 3 + from microprobeanalysis is possible. Fetot/Mg is plotted in Fig. 7, which shows the tendency of

to be lower for Trescolmen and Confin than for the Vals samples. This qualitatively indicatesa lower equilibration temperature for the latter locality, in general agreement with theregional trend of increasing temperature recorded by garnet-clinopyroxene geothermometry.Absolute temperatures derived from Green & Hellman's calibration of K$rlpbe(P,T),however, are about 100 °C higher than those estimated with the garnet-pyroxeneequilibrium (1). This probably indicates that a considerable part of the iron in the Adulaphengites is ferric.

Epidote group minerals (zoisite, clinozoisite, epidote) are widespread as a primary phase ineclogite assemblages throughout the Adula Nappe. Their compositions do not show anysimple regional distribution pattern. Garnet amphibolites in the North usually containFe-rich clinozoisite, whereas kyanite eclogites at Trescolmen and further south more oftencontain Fe-poor zoisite. However, more Fe-rich clinozoisite occurs in Ca-enrichedmetarodingitic eclogites at Gagnone (Evans el al, 1979) and Duria (Heinrich, 1983),indicating that the type of epidote group phase is mainly determined by bulk rockcomposition. In all areas except Arami and Duria, apparently coexisting pairs of zoisite andmore Fe-rich clinozoisite have been observed. Eclogites at Arami contain zoisite as the onlyprimary epidote group mineral.

Fe3 + Al_, exchange in epidote group minerals is roughly correlated with that inassociated omphacites, as calculated from stoichiometric constraints, giving some credence


to the Fe2 + /Fe 3 + estimation procedure used for omphacite. No such correlation has beenfound for stoichiometrically estimated Fe 2 + and Fe 3 + contents of the eclogitic amphiboleswhich are therefore considered to be very uncertain.

1.0

0.5

0.0

log

0• A

-

/

(Fe/Mg)9ar

ValsConfin

Trescolmen >

/

'A

A'/-

A/ :_

log (Fe/Mg)phe

-1.0 -0.5 0.0

FIG. 7. Logarithmic distribution plot of Fe'^/Mg between garnet and primary phengitic white micas from Vals,Confin and Trescolmen. The open triangle with the highest Fe/Mg value represents the garnet and phengite core of

the eclogite fades metapelite from Trescolmen discussed by Heinrich (1982, table 2, sample Ad85).

Amphibole, as a primary eclogitic phase, shows a wide compositional variation, ranging(after Leake, 1978) from impure actinolite in a metadolomite through hornblende,tschermakitic hornblende and barroisite nearly to end-member Mg-glaucophane. Theamphibole in most of the mafic quartz eclogites is common hornblende with a ratio oftetrahedral Al to the total octahedral occupancy of Al + F e 3 + + C r 3 + + T i 4 + close to one(Fig. 8a). Taken alone, the compositions of the texturally primary eclogitic amphibolesfrom the Adula Nappe are thus not very different from those of high to intermediate pressurefacies series amphibolites (e.g. Laird & Albee, 1981). This observation is in contrast toomphacite assemblages from the Western Alpine Sesia Zone, where amphiboles stronglydisplaced from glaucophane are restricted to quartz-undersaturated assemblages (Koons,19826).

Amphibole composition in the Adula eclogites is clearly correlated with the jadeitecontent of the associated omphacite by the coupled exchange Na(AJ,Fe3+)Ca_ ,(Mg,Fe2+)_,,the jadeite or glaucophane exchange vector (Fig. Sb; Heinrich, 1983, figs. 7.3, 7.4). Bulk rockcomposition is certainly one major control on amphibole and pyroxene composition, but inaddition systematic regional differences in mineral compositions are apparent from tie-linerotations such as those shown in Fig. 9. At a given jadeite content of clinopyroxene, the totalNaAl-content of the coexisting primary amphibole tends to increase going from the northernto the southern Adula Nappe.

This can be explained by a regional shift in the continuous net-transfer reaction

NaAlSi2O6(cpx) = + 3 SiO2 (7)jadatc componentin dinopyroieoe

edenile exchangeID amphibole

quartz

140 C. A. HEINRICH

d.U i

1.5

1.0

0.5

no*

glaucophane

toct3 + 4 +

Fcpx

tremolite

tsch'ermakite

/ / 12-2

amp

edenite

I.O

0.5

0.0

f

(D M51-17

0.5

0.00.0 0.5 1.0

1.0

0.8

0.6

0.4

0.2

1.0

0.8

0.6

0.4

0.2

1.0

0.8

glaucophane

t(Na/Na+Ca)M

Gagnone, Arami

cpx ampCH271

tremoliteE2-2

edemte

Confin.Trescolmen

E5-10

E5-6

M25 ' M4SM51-12

Ad51-17

Vals

Ad42

0.2

1.5 2.0 0.0 0.2 0.4 0.6 0.8 1.0FIG. 8. Mineral composition (in atoms per formula unit, Tables 2 and 3) of primary amphiboles (open symbols) andcoexisting clinopyroxenes (full symbols and black bars) in quartz-beanng mafic rocks. Each symbol outlines the fullcomposition range of between 5 and 20 point analyses per mineral and per sample. Arrows indicate zonation trendsfrom grain core to rim. The positions of some important NCMASH end-members of amphibole are indicated forcomparison, (a, left column) Total octahedral occupancy, O d 3 + 4 + = A l ^ + Fe3"1" + C r 3 + +Ti*+ , versus tetra-hedrally coordinated Al. (fc, right column) Plots of (Na/Na + Ca)*" and (Na/Na + Ca)*" of coexisting amphibolesand clinopyroxenes, respectively, on the vertical axjs (measuring the jadeite or glaucophane exchange vector,Na(AlFe3 + )Ca_,(Mg,Fe?.^), versus NaA in amphibole on the horizontal axis (edenite exchange, NaAlD_,Si_,).

ECLOGITE METAMORPHISM IN THE ADULA NAPPE

A + N,

141

C + 2NAd«1

M2S

Trescolmen

E5-10

Confin

M42VAJO

Vals

FIG. 9 Coexisting amphibolcs (open symbols) and clinopyroxenes (full symbols) in the 'ACF-deluxe diagram ofThompson (1981, fig. 28). The coordinates

A + N = AljO3 + Fe2O3 + Cr2O3 + TiO2 + (Na2O + K2O)C + 2N = CaO + 2(Na2O + K2O)

M = MgO + FeO + MnO

correspond to a projection along the exchange vector NaSiCa_,Al_, (plagioclase vector). Amphiboles from thesouthern Adula Nappe (Gagnone) coexisting with a clinopyroxene of a given jadeite content tend to be moreNaAI-rich than those in the central and northern part of the nappe (notably Confin). In addition to this regionaltrend, a similar tie-line rotation is also recorded by grain zonations in individual eclogite samples (e.g. Ad25-9-3

from Trescolmen, indicated by arrows).

142 C. A. H E I N R I C H

0

FIG. 10. Log-log diagram for the net-transfer equilibrium (7) showing analytical data of coexisting clinopyroxenesand amphiboles from quartz-saturated assemblages. Equilibrium (7) controls the edenite exchange in amphiboles(NaAln_,Si_,, vertical axis) as a function of the jadeite content of the coexisting clinopyroxene (horizontal).Glaucophane Ad61-9-l plots below the diagram because its very low Allel and NaA is not well constrained by theanalysis. The systematic regional differences in KN probably reflect a regional increase in temperature from north to

south recorded by the hydrous eclogite assemblages (see text).

(where • denotes a vacancy in the A-site of the amphibole). The logarithmic mass-actionconstant

a N a(8)

aAlSi,O,

can be approximated for analysed pairs of amphiboles and clinopyroxenes (coexisting withexcess quartz) by

logKN~log 'i —6

[(NaM 2AlM 1)]c p (9)

where the element symbols refer to atoms per formula unit in the crystallographic siteindicated by the superscript (cf. Tables 2, 3). In Fig. 10, numerator us. denominator of thisexpression are plotted logarithmically, with contours of constant KN indicated by thediagonal lines. Alternative ways could be used to express Ks for a given pair of analyses, butthe diagram shows qualitatively the regional shift of equilibrium (7) from more jadeiticclinopyroxenes in the northern and central parts of the Adula Nappe (Confin, Trescolmen) tomore edenitic amphiboles in the south (Gagnone and Arami).

The edenite exchange in amphiboles from Vals (boxes outlined with fine lines) does not


seem to be effectively controlled by the composition of the coexisting clinopyroxene,probably because the latter is fixed near jd5 0 by preferential stabilization due to ordering (seeFig. 36).

Equilibrium (7) is the high-pressure equivalent of the equilibrium

NaAlSi3O8(plg) = NaAln^1Si_1(amp)+4 SiO2 (10)tlbite component edcnilc cichangc quartzID pligiodase in ampbibole

which controls the edenite exchange in amphiboles of quartz-bearing amphibolites atpressures below the breakdown of plagioclase to pyroxene. Laird (1980) and Spear (1981)observed that NaAlD^Si . ! increases in plagioclase-quartz-amphibole assemblages withincreasing metamorphic temperature. From this and the small volume change associatedwith reaction (7) (Au°(7) = - 1 0 8 J bar"1 mole"1; using data from Helgeson et al., 1978) itcan be predicted that the high-pressure equilibrium (7) is essentially temperature-sensitiveonly, shifting to the right with increasing temperature. The systematic regional change in thereaction coefficient XN of reaction (7) indicated by Fig. 10 is therefore in good agreement withthe regional gradient in temperature recorded by Fe/Mg exchange between garnet andclinopyroxene.

Where large hornblende poikiloblasts are compositionally zoned, this always involves anincrease in tetrahedrally coordinated Al and the ratio NaA/NaM4 from core to rim (Fig. 8).This indicates a trend of increasing XN (7) (and therefore probably an increase intemperature) during their growth.

Conclusions from mineral chemistry

Comparison of the compositional variations in texturally 'primary' eclogitic hydrosilicatesand associated garnets and clinopyroxenes has shown:

(a) a general correlation between the compositions of hydrous and anhydrous minerals inrock suites from small areas of the Adula Nappe, indicating bulk rock compositional controlin isofacial, often high-variance assemblages;

(b) regional shifts in continuous solid-solid equilibria involving 'primary' hydrosilicates,which qualitatively agree with regional differences in crystallization temperatures andpressures recorded by mineral equilibria among the anhydrous eclogite minerals;

(c) zoning patterns in 'primary' amphiboles which are compatible with zoning in garnet,both probably recording a path segment of increasing temperature during recrystallizationof individual eclogites.

In support of the textural evidence, the mineral chemistry indicates that the crystallizationof the 'primary' hydrosilicates is closely linked, and grossly coeval, with the crystallization ofthe anhydrous eclogite phases.

PHASE RELATIONS IN HYDROUS QUARTZ ECLOGITES

Analysis of phase relations and accurate deduction of mineral reactions between hydrouseclogitic assemblages requires at least ten system components: SiO2, TiO2, A12O3, Fe2O3,FeO, MgO, CaO, K2O, Na2O, and H2O; neglecting minor MnO, Cr2O3 and trace elements.Quartz and rutile are the only two phases which are in excess in all the studied samples, andwhich therefore allow a thermodynamically valid projection in the sense of Greenwood(1967,1975). For the Adula eclogites, K2O can be neglected, because its amount in the rock isreflected by the amount of potassic white mica alone, if the small exchange of KNa_, in

144 C. A HEINRICH

amphibole and paragonite is neglected. Further reduction of components can be achieved bycombination of components or condensation along exchange vectors (Thompson et al., 1982).Condensation of FeO + MgO corresponds to a projection along the exchange vectorFeMg_ t , but this is a thermodynamically valid projection only if the two components do notpreferentially fractionate between the phases under discussion, i.e. if they in fact behave likea single component. Although approximately true in the case of the FeMg_! exchangebetween clinopyroxene and amphibole, this prerequisite obviously does not hold for garnetwhich is preferentially enriched in Fe and has therefore been excluded from the analysis. Thisomission may be the most serious limitation in a more general application of theSchreinemakers analysis developed below, but it will not affect the main conclusions.

For a graphical representation of the phase relations in the condensed NCMASH modelsystem (Thompson et al., 1982), two additional projections are necessary. The approach usedhere is to project tentatively from H2O and along the Fe3 +Al_ [ vector into the tetrahedronNCMA, where

N = Na2O + K2OC = CaO

M = MgO + FeOA = Al203 + Fe2O3 (Fig. 11).

As a consequence, any reaction relationships apparent from NCMA chemographies ofobserved assemblages (Fig. 12) will have to be discussed in terms of differences intemperature and pressure, as well as the level of the activities (a) or chemical potentials (n) ofH 2O and Fe3 +Al_j.

projection fromqfzH20

CaO

Na2O

MgO + FeOFIG. 11. Perspective representation of primary minerals of Adula quartz eclogjtes in the NCMA tetrahedron. Inaddition to projection from quartz and condensation along the exchange vectors Fe I + Mg_, and Fe3 + Al_,, thetetrahedron involves projection from H2O, despite the indication from phase relations that the activity of H2O isnot externally controlled in the Adula assemblages, as discussed in the text. The internal tetrahedronhbl-epi-alb-omp represents the common amphibolite assemblage in the Vals area of the northern Adula Nappe.Glc + epi (dashed line intersecting the plane hbl + alb + omp) occurs as an alternative assemblage in the same area,

probably due to higher levels of Fe 3 + Al_, in some rock compositions (see text).


Topological constraints from Schreinemakers analysis

Disregarding initially the compositional variations along Fe3 + Al_,, and neglectingcomplications which may arise from miscibility gaps in pyroxenes and possibly other phasesat low temperature, the condensed NCMASH-species hornblende, omphacite, epidote(including zoisite/clinozoisite), paragonite, albite, kyanite, quartz and H2O are related by sixreactions. The stoichiometry of these reactions is given in Table 4 for a set of representativecondensed phase compositions. Stoichiometric coefficients vary somewhat according toamphibole and omphacite compositions. However, their sign, and hence the topology of thededuced Schreinemakers grid (Fig. 12), does not change within the full range of phase com-positions observed in the quartz-bearing assemblages of the Adula Nappe. Among the sixreactions only the degenerate reaction

par + qtz = alb + kya + H2O [omp, hbl, epi] (11)

in the NASH-subsystem has been tested experimentally by Chatterjee (1972) and constrainedat higher pressures by the experiments of Holland (1979b). Due to a lack of thermodynamicdata and the complexity of the amphibole solid solution (Heinrich, 1983, p. 130) the reactionsinvolving hornblende cannot be estimated. Only the sign of their slopes can be realisticallyinferred by combining the principles of Schreinemakers (e.g. Zen, 1966), approximatecalculations of reaction volumes (Table 4), and the assumption that reaction entropies aredominated by H20-liberation and (for the H2O-conserving reaction) by the octahedral totetrahedral coordination change of Al (cf. Holland & Richardson, 1979). Fig. 12 showspossible relative positions of the six reactions as heavy lines in a P-T diagram.

Unless aHjO is fixed externally (e.g. by the presence of a hydrous fluid), only the reaction[H2O] is univariant (and hence truly discontinuous in the 6-component model systemNCMASH), because it involves 7 solid phases. These minerals uniquely fix /IH2O and allmineral compositions at any given point along the line [H2O]. The variation of theseparameters along the line [H2O] is exemplified by its intersection with contours for constantaHlO as fixed by the equilibrium par + qtz = alb + kya + H2O (fine broken lines), and withcontours for constant Xjd in omphacite as fixed by the equilibrium alb = jad + qtz (fine fulllines). Note that the increase of Xid along [H2O] with falling temperature suggested by theintersecting contours is arbitrary; [H2O] could equally well have a steeper slope than thejadeite-contours, in which case the variation would be reversed.

The other five H2O-involving reactions [alb], [kya], [qtz], [par] and [omp, hbl, epi] aredivariant bands even in the NCMASH model system, unless aHl0 is given a specified value.Correspondingly, the point I will shift to lower temperatures along the line [H2O] as aH2O islowered. In Fig. 12 the point I and the H2O-involving NCMASH-reactions are placedarbitrarily to represent positions which would be compatible with aHlO ~ 05.

Comparison with natural assemblages

The derived Schreinemakers topology is in agreement with the phase assemblagesobserved in the four areas Gagnone, Trescolmen, Confin and Vals (shown as the tetrahedralinsets in Fig. 12), as well as with the thermobarometric results (indicated by V, C, T, G;cf. Fig. 6). Except for minor tie-line rotations reflecting regional shifts in continuousreactions, the chemographies for Gagnone and Trescolmen are similar, with kyaniteoccurring in sufficiently aluminous rocks. An additional epidote phase (usually zoisite) iscommon in both areas in relatively calcic, typically mafic rock compositions, coexisting withan aluminous hornblende. In more sodic compositions, kyanite + omphacite + quartzcoexist with paragonite and glaucophane in the Trescolmen area. The latter minerals have

146 C. A H E I N R I C H

TABLE 4

Reactions between hydrous and anhydrous phases of mafic quartz eclogites in the condensedNCMASH system

Phase

qtzkyaalbompepiparhblV

NCMASH-condensed composition

SiO2

Al2SiO2

NaAlSi3O, (low-alb)Na 3 Ca 3 Al 3 Mg 3 Si 2 O 6

Ca2Al3Si3O12(OH)(zoi)NaAl3Si3O10(OH)2

• 3NaCa, 3Mg3 5AI3Si6 ,O22(OH)2

H2O

estimated volume change Av" (J bar"1

(Jbar~')

2 27441

10016 32

13 59132026-91

1-85

)

Reference

(R)(R)(R)(R)(He)(Ho)(He,K)(DH)

Sloichiometrlc coefficients

[omp)

Lepi]

_ 1

+ 1+ 1

- 1

+ 10-80

[kya]

-6

+ 10- 1 4+ 2- 5+ 2+ 2

+ 16-7

[W2O]

+ 4+ 2- 8

+ 14- 2+ 3- 2

- 1 5 1

In reactions (Fig

[alb]

- 4+ 10

+ 14- 2- 5- 2+ 8- 8 6

[par]

+ 1+ 5- 5

+ 14- 2

- 2+ 3

-127

12)

[««]

+ 6- 4

+ 14- 2- 1

- 2+ 4

- 1 1 9

Omphacite composition in all Adula rocks containing a calcic (hornblende, epidote) plus a sodic/aluminous phase (albite, kyanite,paragonite) is close to jadeite30diopside30. Amphibole has been idealized to a hornblende formula hbl =0-5 tremolite + 0-5pargasite + 0-5 NaAICa_ ,Mg_,+0-5 Al2Mg_,Si_!, which closely approaches the condensed composition of hornblendes co-existing with jadeite50-omphacites.

Molar volumes have been estimated using data from (He) = Helgeson et al. (1978), (R) — Robie et at (1979), (Ho) = Holland(1979b) and (K) = Koons (1982a), neglecting non-ideal mixing in omphacite and amphibole. The volume for H2O is taken fromDH = Delany & Helgeson (1978) at the somewhat arbitrary conditions of 12 kb/700°C, but variation of p°(H2O) within the range ofequilibration conditions of the Adula Nappe has only a small influence on AD° of the reactions

not been found in Gagnone, although they would be compatible with the observed topology,and also with the estimated P~T conditions and the experimental stability of glaucophane +paragonite (Koons, 1982a). Their absence is more likely a consequence of selectiveoverprinting of these assemblages during subsequent amphibolite facies metamorphism(Heinrich, 1982), rather than a reflection of the differences in original eclogite facies P-Tconditions between the two areas.

Among the assemblages from Confin, a reaction relationship

par + epi + hbl+qtz = omp + kya + H2O [alb] (12)(Ad48-9-5) (E5-10)

is evident from tetrahedron ' C (Fig. 12), which must involve H2O to be balanced. Since thetwo incompatible assemblages occur in the same eclogite lens, sharing a pronounced minerallineation, crystallization under different conditions of P and T is unlikely. Amphibole andomphacite in the two samples have similar and low total Fe-contents, and therefore bulkcompositional differences with regard to Fe2 + Mg_L and Fe3 + Al_] are an equally unlikelyexplanation for the different assemblages. Most probably the two assemblages indicate localdifferences in the activity of H2O. The assemblages from Confin are represented in Fig. 12 bythe Schreinemakers sector [alb, H2O], whereby the assemblage omp + kya is stabilizedrelative to par + epi + hbl + qtz by a shift of reaction [alb] due to a lower level of aHiO.

The northern Adula Nappe, including Vals, is known for its occurrences of mafic rockswith glaucophane. Throughout this area, glaucophane is invariably associated with Fe-rich,true epidote (e.g. Ad65-0-8, Table 3; Gansser, 1937; Plas, 1959; Oberhansli, 1978). Whereverpresent, coexisting omphacite is distinctly apple green in thin section indicating a highacmite content (Table 2). The coexistence of glaucophane and a pure CaAl-phase like zoisiteor lawsonite, considered diagnostic of the blueschist facies and common in mafics from the

E C L O G I T E M E T A M O R P H I S M IN T H E A D U L A N A P P E 147

FIG. 12. Partly schematic P -T diagram showing the relative position of some equilibria among NCMASH-condensed minerals of quartz-bearing hydrous mafic rocks (heavy lines; see Table 4 for stoichiometry), incomparison with observed mineral assemblages (tetrahedral chemographies in the projection of Fig. 11) and theirapproximate crystallization conditions derived from mineral compositions for the localities Vals (V), Confin (Q,Trescolmen (T) and Gagnone (G) (Fig. 6). Full fine lines = contours of constant jadeite content (100 Xjd) for allassemblages containing albite + omphacite; equilibrium (4). Broken lines = constant aHlO in assemblages containingparagonite + albite + kyanite +quartz; equilibrium (11) Stippled lines = contours for constant product (100X^ • alllO) for assemblages containing par + omp + kya; equilibrium (5). H2O activites refer to a standard state with

flHio = ' f° r t n e P u r e aqueous fluid, irrespective of the physical state of the component H2O.

Sesia zone (Western Alps; Koons, 1982b) or in the Franciscan terrain (California; Brown &Bradshaw, 1979), has never been observed in the Adula Nappe.

Glaucophane-epidote rocks in Vals are relatively subordinate to the more common(±omphacite-bearing) garnet-epidote-albite amphibolites in this area. In condensedNCM ASH space, the glaucophane and amphibolite assemblages appear to overlap (Fig. 11),linked by the reaction

glc + epi = alb + omp + hbl + qtz + H2O. (13)

It is possible that this equation again reflects different values of H2O activity between

148 C. A HEINRICH

cofacial rocks. However, Fe3 + Al_! strongly fractionates between these minerals, with thesequence

FeAL,(epi) » FeAl_,(hbl) ^ FeAl.^omp) ~ FeAl_,(glc) » FeAl-^alb)

estimated from microprobe analyses. Using non-condensed mineral compositions from therepresentative samples Ad42-9-14 and Ad65-0-8 (Tables 2, 3), a numerical test of reaction(13) using the method of Greenwood (1967,1975) has shown that the two assemblages do nothave overlapping phase polyhedra in multicomponent space also including Fe2O3.Glaucophane-epidote assemblages in northern Adula Nappe thus reflect relatively Fe3 + -rich bulk compositions, and therefore not necessarily different P-T-aHlO conditions,compared to the more common epidote-albite amphibolite assemblages in the same area.Glaucophane ( + epidote) assemblages are not considered explicitly in the followingdiscussion, but the inclusion of this amphibole with 'hbP would not alter the topology ofFig. 12.

In contrast, the association paragonite + omphacite found in the impure dolomiteAd41-9-20 from Vals is related to the more common epidote-amphibolite assemblages inthis area by

omp + par + qtz = alb + hbl+epi + H2O [kya].(Ad41-9-20) (Ad42-9-14) " V

This relation was also tested numerically, including garnet approximating the analysedcompositions. The result showed a small amount of garnet to be involved on the left side ofthe reaction, but the sign of H2O and the other solid did not change. Since all phases havevery similar compositions in the two samples Ad41-9-20 and Ad42-9-14, the relation (14)indicates local differences in aHlO as the most likely cause for the different mineralassemblages. Note that the typical assemblage of the carbonate-poor or carbonate-freeamphibolites from Vals indicates the lower level of H2O activity, whereas the dolomite- andcalcite-rich schist Ad41-9-20 contains the assemblage indicating the higher H2O activity.This is in agreement with the interpretation discussed below, that a mechanism other thanfluid dilution by CO2 was the main cause for local variations in /iH2O during high-pressuremetamorphism of the northern Adula Nappe.

In the Schreinemakers grid of Fig. 12, the Vals assemblages are compatible with the sector[kya, H2O], provided that the less hydrous assemblage alb + epi + hbl is stabilized relative toomp + par by a shift of reaction [kya] (14) due to a lower level of H2O activity.

It is interesting to conclude that among probably isofacial samples from Vals, which havea mineralogical range transitional between epidote amphibolite and eclogite facies, the'eclogitic' assemblage omphacite + paragonite ( +garnet) is stabilized, not destabilized asperhaps expected, by increased H2O activity relative to the typical amphibolite assemblagealb + hbl + epi. A similar conclusion has been reached by Brown & Bradshaw (1979) foromphacite in hydrous rocks transitional between lawsonite-blueschist and low-temperatureeclogite facies in California.

Isograds of high-pressure metamorphism

The general compatibility of the observed phase topologies with the theoreticalSchreinemakers analysis and the thermobarometric results shows that the reactions [alb]and [H2O] have the character of isograds of high-pressure regional metamorphism. The factthat these 'isograds' are not separated by sharply defined geographic boundaries is a directconsequence of the variations in water activity indicated by isofacial assemblages from Vals

E C L O G U E METAMORPHISM IN THE ADULA NAPPE 149

and Confin, but it probably also reflects the low sampling density and the complications bypost-eclogite fades deformation. The following interpretation of the geological meaning ofthese 'isograds' and the physical significance of the reduced water activities in Adulahigh-pressure metamorphism is therefore tentative.

FLUID PRESENCE AND FLUID ABSENCE IN ADULAHIGH-PRESSURE METAMORPHISM

Metamorphism is often assumed to be a process involving the ubiquitous presence of a freephase. In prindple, local and large-scale differences in fluid composition could be invoked toexplain the local and regional differences in eclogite mineralogy observed in the AdulaNappe. However, the combined geological, textural, and mineralogical information isexplained more easily by a process involving a regional change from fluid-absent tofluid-present high-pressure metamorphism, in response to different P-T paths experiencedby different parts of the present Adula Nappe.

The majority of the mafic high-pressure assemblages in the northern part of the AdulaNappe (Vals to Confin) formed from high-grade amphibolites, i.e. from hydrous protolithswhich had already been dehydrated to some extent by Pre-Alpine metamorphism. As aconsequence, high-pressure metamorphism was probably dominated by H2O-conservingreactions forming garnet + omphacite-bearing epidote-albite amphibolites and hydrate-richeclogites. Their high-pressure mineralogy was in part determined by the bulk hydratecontent of the original protolith, in the way generally discussed by Thompson (1955, p. 99).Large-scale differences in recorded P-T conditions within this region of H2O-conservingmetamorphism are now reflected by the geographic distribution of mineral assemblageswhich are regionally separated by a H2O-conserving 'isograd' reaction:

alb + epi + hbl = omp + kya + par + qtz [H2O]. (15)

As a consequence of largely H2O-conserving mineral reactions among H2O-undersaturatedassemblages, high-pressure metamorphism for the majority of the mafic rocks in thenorthern Adula nappe probably occurred without saturation of the grain boundaries with afree fluid phase. The predominance offluid-absent metamorphism (in the sense of Thompson,1983), inferred for the northern Adula Nappe, not only explains aH2O gradients over smalldistances, but is compatible with the often incomplete recrystallization and generally finegrain size of most high-pressure assemblages in this area. Fluid presence during high-pressure metamorphism of the northern Adula Nappe was probably restricted to thesubordinate Mesozoic metasediments (such as the paragonite- and omphacite-bearingcarbonate schist Ad41-9-20, see above) which had not been metamorphosed previously, andto localized veins crosscutting some of the older rocks.

In contrast to the water-conserving isograd within the northern Adula Nappe, thereaction separating the phase topology of Confin from that of the southern Adula Nappe (e.g.Trescolmen, Gagnone) involves major dehydration:

par + epi + hbl + qtz = omp + kya + H 2 0 . (12)

This dehydration 'isograd' is interpreted to mark the transition from an area of dominantlyH2O-conserving metamorphism in the north to an area further south, where a similar suiteof hydrous mafic protoliths reacted by prograde dehydration metamorphism to kyaniteeclogites, which still contain some primary hydrosilicates but in lower modal abundance. Theabsence of mineralogical indications for local differences in aH3O between cofadal eclogitesin the southern Adula Nappe indicates that reaction (12) also marks the regional transition

150 C. A HEINRICH

from largely water-undersaturated metamorphism to a metamorphic process during whichaHlO was controlled externally by the rate of dissipation of water from the site of itsproduction (Thompson, 1955, p. 96). Eclogite formation in the southern Adula Nappeprobably involved the local and temporary presence of an H2O-bearing fluid phase.

Local coexistence of the product assemblage of reaction (12) with a macroscopic fluid isshown by quartz veins containing omphacite + kyanite. These veins are particularly frequentat Trescolmen, where also the coarsest-grained eclogites in the Adula Nappe are found. Theabsence of any wall rock alteration adjacent to these veins, and the presence of all matrixminerals of the kyanite eclogites immediately adjacent to the vein walls are compatible witha stage of fluid presence not only in the veins, but probably also in the mafic eclogitesthemselves. The veins are interpreted as hydraulic fractures representing the collectingchannelways of escaping volatiles produced during eclogite formation (cf. Walther & Orville,1982). From this it would be expected that the fluids forming the veins at Trescolmen wereH2O-rich, but no primary fluid inclusions have been identified which would allow a directestimation of the composition of the vein-forming fluid.

Outside the mafic eclogite lenses, fluid-present metamorphism was probably notregionally extensive, even during dehydration of the high-grade mafic eclogites. If the geneticinterpretation of a Variscan high-grade pre-metamorphosed basement unit being subjectedto subduction metamorphism is correct, then any water produced by eclogitization of maficamphibolite layers was probably consumed by paragonite- and phengite-forming hydrationreactions in the adjacent granites and metapelites. Although texturally not preserved, thesehydration reactions may have been similar (but opposite in sign) to the dehydrationreactions which occurred during amphibolite facies overprinting of the country rocksenclosing the mafic eclogites (Heinrich, 1982). Because of the quantitative predominance ofthe pelitic, quartzofeldspathic and granitic country rocks relative to the mafics, the amountof fluid produced by high-pressure dehydration of the mafic rocks is unlikely to have beensufficient to saturate the country rock assemblages and to cause regionally extensive fluidpresence during eclogite facies metamorphism. As a consequence, the net loss of volatilesduring high-pressure metamorphism from the Adula Nappe as a whole was probably atmost minor.

Even in the mafic eclogites of the southern Adula Nappe, fluid-present high-pressuremetamorphism was probably restricted to times when major dehydration reactions were inprogress. As the important dehydration reactions are sensitive to increasing pressure, the lastevent causing fluid saturation in the kyanite eclogites may have occurred closer to their peakmetamorphic pressure, rather than near the peak temperature reached by the same rockslater on their P-T path.

CONCLUSIONS

The zoisite/epidote- and hornblende-bearing kyanite eclogites of the southern AdulaNappe are a characteristic example of'common' or 'type-B' eclogites (Smulikowski, 1964;Coleman et al., 1965). The petrological data presented above, in conjunction with thosediscussed by Heinrich (1982), lead to the conclusion that type-B quartz eclogites can formin situ (i.e. with a metamorphic history shared by their gneissic country rocks) by progradedehydration metamorphism from more hydrous protoliths. It should be stressed that theincompatible metamorphic pressures often recorded by mafic type-B eclogites and adjacentplagioclase-rich amphibolite-facies gneisses do not provide unequivocal evidence againsta common regional high-pressure metamorphic history shared by the two rock types(Heinrich, 1982).

ECLOGUE METAMORPHISM IN THE ADULA NAPPE 151

Prograde dehydration of mafic rocks to quartz-kyanite eclogite is confined to depthsbelow the normal continental Moho, and requires subduction or tectonic crustal thickening.Field data suggest that the transition near 12-15 kb/500-600 °C from eclogite rich inparagonite + epidote + amphibole to quartz-kyanite eclogite according to reaction (12)represents an important dehydration step in progressive high-pressure metamorphism ofmafic rocks. However, even at considerably higher P-T conditions up to at least 20 kb/800 °C, minor zoisite, hornblende and phengite persist as stable hydrosilicates inquartz-kyanite eclogites of suitable bulk composition.

Near 12 kb/500 °C or less, garnet, omphacite and quartz can coexist stably with allminerals typical of albite-epidote amphibolites, but the assemblage albite + epidote +hornblende + quartz + omphacite ( +garnet) requires fluid-absent metamorphism or dilu-tion of any hydrous fluid by other components. Rather than omphacite-garnet amphibolite,the association of omphacite + quartz ( +garnet, amphibole, epidote) with paragonite is themost likely 'eclogitic' assemblage to form by prograde dehydration metamorphism at crustalpressures of less than 10 kb.

ACKNOWLEDGEMENTSThis research was carried out as part of a Ph.D. project at ETH Zurich. I have greatly

benefited from working with V. Trommsdorff and A. B. Thompson who supervised theproject, contributing enthusiasm and ideas at all stages of the work through to thecompletion of this manuscript. The petrographic and analytical work relied on the excellentsample preparation by A. Willi, and the help of J. Sommerauer and S. Girsperger with the useof microprobe and computer facilities. I enjoyed many stimulating discussions with P. O.Koons, and with other colleagues at ETH and other Swiss universities. A. S. Andrew,M. Frey, B. Hensen, P. O. Koons and D. J. Whitford read the manuscript at various stages,and their comments were very helpful to improve its content and presentation. Reviews byW. V. Maresch and T. J. B. Holland are gratefully acknowledged. A great effort was providedby the typing office at CSIRO, Institute of Energy and Earth Resources in Sydney, where thispaper was completed. To all of these people, 1 would like to express my thanks for their helpand interest! Financial support by the Schweizerischer Nationalfonds (Project 2.805-0.80) isacknowledged.

REFERENCESAnderson, R. N., Dc Long, S. F., & Schwarz, W. M., 1978. Geophysical and geochemical constraints at converging

plate boundaries. J. Geoi. 86, 731-9.Aurisicchio, C, Bocchio, R., Liborio, G., & Mottana, A, 1982. Petrogenesis of the eclogites from Soazza,

Switzerland. Abstr. 1st Int. Eclogite Conference, Terra cognila, 2, 305.Banno, S-, & Matsui, Y., 1965. Eclogite types and partition of Mg, Fe, and Mn between clinopyroxene and garnet.

Proc. Japan. Acad. 41, 716-21.Baumgartner, L., 1981. Petrologie der Alp dc Confin, Misox GR. Unpublished Diploma thesis. University of Basel.

& Low, S., 1983. Deformation und Metamorphose der Adula-Decke sudwestlich San Bernardino, Schweiz.Schweiz. miner, petrogr. Mitt. 63, 215-32.

Bcarth, P., 1959. Ueber Eklogite, Glaukophanschiefer und metamorphe Pillowlaven. Ibid. 39, 267-86.Black, P. M., 1977. Regional high-pressure metamorphism in New Caledonia: phase equilibria in the Ouegoa

District. Tectonophysics, 43, 89-107.Brown, E. H., & Bradshaw, J. Y., 1979. Phase relations of pyroxene and amphibole in greenstone, blueschist and

eclogite of the Franciscan Complex, California. Contr. Miner. Petrol. 71, 67-83.Buletti, M., 1983. Zur Geochemie und Entstehungsgeschichte der Granat-Amphibolite des Grambarognogebietes,

Ticino. Schweiz. miner, petrogr. Mitt. 63, 233-47.Carpenter, M. A., 1980. Mechanisms of exsolution in sodic pyroxene. Contr. Miner. Petrol. 71, 289-300.Chatterjee, N. D-, 1972. The upper stability limit of the assemblage paragonite + quartz and its natural occurrences.

Ibid. 34, 288-303.& Johannes, W. S., 1974. Thermal stability and standard thermodynamic properties of synthetic 2M1-muscovite, KAl3Si3O10(OH)2. Ibid. 48, 89-114.

152 C. A. H E I N R I C H

Clark, J R., & Papike, J. J., 1968. Crystal-chemical characterization of omphacites. Am Miner. 53, 840-68.Codoni, A. G , 1981. Geologia e petrografia del Pizzo di Claro PhD thesis, Zurich University.Coleman, R. G., Lee, D. E., Beatty, L. B., & Brannock, W. W., 1965. Eclogites and eclogites: their differences and

similarities. Bull. Geol. Soc. Am 76, 483-508.Delany, J. M., & Helgeson, H. C, 1978 Calculation of the thermodynamic consequences of dehydration in

subducting oceanic crust to 100 kb and > 800 °C Am. J. Sci. 278, 638-86.Egli, W., 1966. Geologisch-petrographische Untersuchungen in der NW-Aduladecke und in der Sojaschuppe

(Bleniotal, Kanton Tessin). PhD thesis, ETH Zurich.Ellis, D. J., & Green, D. H., 1979. An experimental study of the effect of Ca upon garnet-clinopyroxene Fe-Mg

exchange equilibria. Contr Miner. Petrol. 71, 13-22.England, P. C, & Thompson, A. B., 1984. Pressure-temperature-time paths of regional metamorphism I Heat

transfer during the evolution of regions of thickened continental crust. J Petrology, 25, 894-928.Ernst, W. G., 1977. Mineralogic study of eclogitic rocks from Alpe Arami, Lepontine Alps, Southern Switzerland

Ibid. 18, 371-98.1978. Petrochemical study of lherzolitic rocks from the Western Alps. Ibid 19, 341-92.& Dal Piaz, G. V., 1978. Mineral parageneses of eclogitic rocks and related mafic schists of the Piemonteophiolite nappe, Breuil-St. Jacques area, Italian Western Alps. Am. Miner. 63, 621-40.

Essene, E. S., & Fyfe, W. S., 1967. Omphacitc in Californian metamorphic rocks. Conlr. Miner. Petrol. 15,1-23.

Evans, B. W., & Trommsdorff, V, 1978. Petrogenesis of garnet lherzolite, Cima di Gagnone, Lepontine Alps. Earthplanet. Sci. Lett. 40, 333-48

& Goles, G., 1981. Geochemistry of high-grade eclogites and metarodingites from the Central Alps. Contr.Miner. Petrol. 76, 301-11.

& Richter, W., 1979. Petrology of an eclogite-metarodingite suite at Cima di Gagnone, Ticino,Switzerland. Am. Miner. 64, 15-31.

Frey, M., Jager, E., & Niggli, E., 1976. Gesteinsmetamorphose im Bereich der Geotraverse Basel-Chiasso. Schweiz.miner, petrogr. Mitt. 56, 649-59.

Fry, N., & Fyfe, W. S., 1969. Eclogites and water pressure. Contr. Miner. Petrol. 24, 1-6.Gansser, A., 1937. Der Nordrand der Tambodecke. Schweiz. miner, petrogr. Mitt. 17, 291-523.Gasparik, T., & Lindsley, D. H., 1980. Phase equilibria at high pressures of pyroxenes containing monovalent and

tnvalent ions. In: Prewitt, C. T. (ed.) Pyroxenes. MSA Rev. Mineralogy, 7, 309-39.Green, D. H., & Ringwood, A. E., 1968. Genesis of the calc-alkaline igneous rock suite Contr. Miner Petrol 18,

105-62.Green, T. H., & Hellman, P. L., 1982. Fe-Mg partitioning between coexisting garnet and phengite at high pressure,

and comments on a garnet-phengite geothermometer. Lithos, 15, 253-66.Greenwood, H. J., 1967. The N-dimensional tie-line problem. Geochim. cosmochim. Acta, 31, 465-90.

1975. Thermodynamically valid projections of extensive phase relationships. Am. Miner. 60, 1-8.Griffin, W. L., Austrheim, A., Brastad, K., Bryhni, I., Krill, A., Krogh, E., Mork, M. B. E., Qvale, H., &

Torudbakken, B., 1985. High-pressure metamorphism in the Scandinavian Caledonides. In: Gee, D. G., &Sturt, B. A. (eds.) The Caledonian Orogen, Scandinavia and Related Areas. Wiley, in press.

Hanny, R., Grauert, B., & Soptrajanova, G , 1975. Paleozoic migmatites affected by high-grade Tertiarymetamorphism in the Central Alps (Valle Bodengo, Italy). Contr. Miner. Petrol. 51, 173-96.

Harley, S. L., & Green, D. H., 1982. Garnet-orthopyroxene barometry for granulites and peridotites. Nature, 300,697-701.

Heinrich, C. A., 1982. Kyanite-eclogite to amphibolite fades evolution of hydrous mafic and pelitic rocks, AdulaNappe, Central Alps. Contr. Miner. Petrol. 81, 30-8.1983. Die rcgionale Hochdruckmetamorphose der Adula-Decke, Zentralalpen (Schweiz). PhD thesis 7282,ETH Zurich.

Helgeson, H. C, Delany, J. M., Ncsbitt, H. W., & Bird, D. G, 1978. Summary and critique of the thermodynamicproperties of rock forming minerals. Am. J. Sci. 278A, 1-229.

Holdaway, M. J., 1971. Stability of andalusite and the aluminum silicate phase diagram. Ibid. 271, 97-131.Holland, T. J. B., 1979a. High water activities in the generation of high pressure kyanite eclogites of the Tauern

Window, Austria. J. Geol. 87, 1-27.1979b. Experimental determination of the reaction paragonite =-jadeite + kyanite + H2O, and internally

consistent thermodynamic data for part of the system Na2O-AI2O3-SiO2-H2O, with applications to eclogitesand blueschists. Contr. Miner. Petrol. 68, 293-301.1983. The experimental determination of activities in disordered and short-range ordered jadeitic pyroxenes.Ibid. 82, 214-20.& Richardson, S. W., 1979. Amphibole zonation in metabasites as a guide to the evolution of metamorphicconditions. Ibid. 70, 143-8.

Jenny, H , Frischknecht, G., & Kopp, J., 1923. Geologie der Adula. Beitr. Geol. Kane Schweiz N.F. 51.Koons, P. O , 1982a. An experimental investigation of the behaviour of amphibole in the system Na 2O-MgO-

Al2O3-SiO2-H2O at high pressures. Contr. Miner. Petrol. 79, 258-67.19826. An investigation of experimental and natural high-pressure assemblages from the Scsia Zone, Western

Alps. PhD thesis 7169, ETH Zurich.

E C L O G U E M E T A M O R P H I S M IN T H E ADULA N A P P E 153

1984. Implications to garnet-clinopyroxene geothermometry of non-ideal solid solution in jadeitic pyroxenes.Contr. Miner Petrol. 88, 340-7.

Krogh, E. J., 1977. Evidence of Precambrian continent-continent collision in Western Norway. Nature, 267, 17-19.& Raheim, A., 1978. Temperature and pressure dependence of Fe-Mg-partitioning between garnet andphengite, with particular reference to eclogues. Contr. Miner. Petrol. 66, 75-80.

Laird, J., 1980 Phase equilibria in mafic schist from Vermont. J. Petrology, 21, 1-37.& Albee, A. L, 1981. Pressure, temperature and time indicators in mafic schist: their application toreconstructing the polymetamorphic history of Vermont. Am. J. Sci. 281, 127-75.

Lambert, I. B., & Wyllie, P. H., 1972. Melting of gabbro (quartz eclogite) with excess water to 35 kilobars, withgeological applications. J. Geol 80, 693-708.

Leake, B. E., 1978. Nomenclature of amphiboles. Am. Miner. 63, 1023-52.Lovering, J. F., & White, A. J. R., 1969. Granulitic and eclogitic inclusions from basic pipes at Delegate, Australia.

Contr. Miner. Petrol. 21, 9-52.Maresch, W., 1983. Die Bildung von Eklogit aus Amphibolit bei progressiver Hochdruckmetamorphose. Fortschr.

Miner. 60, Beih. 1, 137-8.Miyashiro, A., 1973. Metamorphism and Metamorphic Belts. London: Allen & Unwin.Newton, R. C, & Fyfe, W S., 1976. High pressure metamorphism. In: Bailey, D K., & Macdonald, R. (eds.) The

Evolution of the Crystalline Rocks, 101-86.Oberhansli, R., 1978 Chemische Untersuchungen an Glaukophan-fuhrenden basischen Gesteinen aus den

Bundnerschiefern Graubundens. Schweiz. miner, petrogr. Mitt. 56, 139-56.Plas, L. van der, 1959. Petrology of the northern Adula region, Switzerland. Leidse Geol. Meded. 24, 415-602.Robie, R. A, Hemingway, B. S, & Fisher, J. R., 1979. Thermodynamic properties of minerals and related

substances at 298-15 K and 1 bar (105 pascals) pressure and at high temperatures. Bull. US. geol. Surv. 1452.Rubie, D. C, 1984. A thermal-tectonic model for high-pressure metamorphism and deformation in the Sesia Zone,

Western Alps. J. Geol 92, 21-36.Smulikowski, K., 1964. An attempt at eclogite classification Bull. Acad. Polon. Sci. (Sect. Sci. geol. geogr.) 12, 27-33.Spear, F. S., 1981. Amphibole-plagioclase phase equilibria: an empirical model for the relation albite + tremolite =

edenite +quartz. Contr. Miner. Petrol. 77, 355-64.Takasu, A., 1984. Prograde and retrograde eclogites in the Sambagawa metamorphic belt, Besshi District, Japan. J

Petrology, 25, 619^3.Thompson, A. B., 1983. Fluid-absent metamorphism J. geol. Soc. Land. 140, 533-47

& Tracy, R. J., 1979. Model systems for anatexis of pelitic rocks. II. Fades series melting and reactions in thesystem CaO-KAlO2-NaAlO2-AI IO3-SiO2-H2O. Contr. Miner. Petrol. 70, 429-38.

Thompson, J. B., 1955. The thermodynamic basis for the mineral facies concept. Am. J. Sci. 253, 65-103.1981. An introduction to the mineralogy and petrology of the biopyriboles. In: Veblen, D. R. (ed.) Amphiboles

and other hydrous pyriboles—mineralogy. MSA Rev. Mineralogy, 9A, 141-88.Laird, J_, & Thompson, A. B., 1982. Reactions in amphibolite, greenschist and blueschist. J. Petrology, 23,1-27.

Walther, J. V., & Orville, P M , 1982. Volatile production and transport in regional metamorphism. Contr. Miner.Petrol. 79, 252-7

Winkler, H. G. F., 1974. Pelrogenesis of Metamorphic Rocks. Springer, 4th edWyllie, P. J., 1971. The Dynamic Earth. New York: Wiley.Zen, E-An, 1966 Construction of pressure-temperature diagrams for multicomponent systems after the method of

Schreinemakers—a geometric approach. Bull. U.S. geol. Surv. 1225, 1-56.

APPENDIX 1

Sampling locations (referring to Swiss coordinates) and some petrographic details of the rocksstudied by microprobe (cf. Table 1, text). All samples are stored at ETH Zurich, including somecollected by W. Richter, V. TrommsdorlT and R. Oberhansli.

ValsAd42-9-14 (734700/159000) Banded omp-gar amphibolite from Peil, earlier described by Plas (1959,

p. 502). Interlocked hbl+alb poikiloblasts, other minerals form fine-grained matrix and inclusions inhbl and alb. Apparently stable coexistence of omp + alb + qtz.

VA30 (same locality) Layer of fine-grained granoblastic to schistose hydrosilicate-rich eclogite inamphibolite Ad42-9-14; mineral compositions almost identical to Ad42-9-14.

Ad41-9-20 (735500/160400) cm-thick, probably tuflaceous, layer in calc-mica schist from Mesozoiccover of Adula Nappe.

Ad65-0-8 (732750/161300) Garnet glaucophanite with Fe-rich epi and acmitic omp. Equigranularschistose texture.

ConfinAd48-9-5 (731900/147000) Equigranular matrix with linear fabric, enclosing small euhedral gar

porphyroblasts.

154 C A. HEINRICH

E5-10 (same locality) Same texture. Barroisite + kya enriched in 2 cm boudin grading intosurrounding par-rich eclogite.

E5-6 (same locality) Similar texture except poikiloblastic hbl with 'myrmekitic' qtz precipitates. Hblcomposition probably not primary.

Ad51-9-12 (732500/146700) Calcic eclogite rich in apparently coexisting zoi + clz. At contact oftremolite-antigonte schist—possibly metarodingite.

Ad51-9-17 (732500/146700) Dolomite marble of unknown age at serpentinite contact. Poikiloblastsof jadeite-poor omp +primary act+zoi. Partial later overprinting of zoi + dol to chlorite + clc.

TrescolmenAd25-9-3 (733690/139810) Granoblastic matrix with 1 -5 mm kya and hbl poikiloblasts. Omp-kya-

qtz veins in same outcrop.Ad61-9-1 (733460/139690; block) Nematoblastic very qtz-rich eclogite with nearly Ca-free Mg-

glaucophane. Talc later, possibly pseudomorphing earlier carbonate (cf. Holland, 1979a, p. 14).

GagnoneCH271 (707960/130750) Texture as Ad25-9-3. Omp-kya-qtz veins occur in same lens.Mgl63K (708360/131060) Banded granoblastic matrix with gar as euhedral porphyroblasts and

atoll-crystals. Contact of chlorite-enstatite metaperidotite (see Evans el at., 1979; including analyses).Mgl63-4-8 (same locality) Equigranular layer rich in hbl + zoi, with strong grain alignment shared

by omp and hydrosilicates.E2-2 (708090/130860) Partly overprinted eclogite with porpyroblastic garnet. Clz and zoi show no

reaction relationship. Some hbl secondary.

AramiMg9-5-12c (719000/121000) Equigranular eclogite from mafic envelope of garnet lherzolite lens.

Weak schistosity defined by omp and zoi. All minerals including gar strikingly homogeneous. Hbl inmatrix and as gar-inclusions, with same composition.

Documents

Eclogite Fades Regional Metamorphism of Hydrous Mafic Rocks … · 2017-08-02 · Eclogite Fades Regional Metamorphism of Hydrous Mafic Rocks in the Central Alpine Adula Nappe by